Secondary Logo

Can Any Metabolites Partially Alleviate Fatigue Manifestations at the Cross-Bridge?


Medicine & Science in Sports & Exercise: January 2004 - Volume 36 - Issue 1 - p 20-27
doi: 10.1249/01.MSS.0000106200.02230.E6
CLINICAL SCIENCES: Symposium—Protecting Muscle ATP during Fatigue: How Metabolites Contribute to Crisis Management

MYBURGH, K. H. Can Any Metabolites Partially Alleviate Fatigue Manifestations at the Cross-Bridge? Med. Sci. Sports Exerc., Vol. 36, No. 1, pp. 20–27, 2004. During exercise, intracellular homeostasis depends on the matching of adenosine triphosphate (ATP) supply and ATP demand. Metabolites play a useful role in communicating the extent of ATP demand to the metabolic supply pathways. During fatigue from high-intensity exercise, a major change in the intracellular milieu of skeletal muscle is not ATP depletion but metabolite accumulation that affects the actomyosin cross-bridge interaction. The resulting reduction in myosin ATPase activity, cross-bridge turnover rate, and velocity of contraction can be considered a useful downregulation of ATP demand. Although maximal force is reduced, it is reduced less than myosin ATPase activity. In combination, efficiency of force production at the cross-bridge is thus enhanced. This is a second useful role for metabolites during fatigue because the total ATP cost per unit of force is partially reduced. Theoretical models predict that ADP may alleviate some effects of fatigue by further enhancing cross-bridge efficiency, thus providing a third useful role for metabolite accumulation. Recent experimental evidence reviewed here suggests that this occurs when ATP concentration is dramatically reduced. Single-fiber chemical analyses of fatigued muscle show lower ATP concentrations than other methods, but whether the appropriate microenvironments for effective competition by ADP for the nucleotide binding site occur around some or all of the cross-bridges remains technically difficult to prove at this time. During fatigue, muscle activation is also decreased, a response that potentially has the greatest effect on ATP demand-supply matching. I propose that the mismatch between the expected force production relative to muscle activation and the reduced force production from inorganic phosphate accumulation is the peripheral signal for reduced activation and is therefore the fourth useful role of metabolites in alleviating fatigue.

Department of Physiological Sciences, University of Stellenbosch, Stellenbosch, SOUTH AFRICA

Address for correspondence: Kathryn H. Myburgh, Department of Physiological Sciences, University of Stellenbosch, Private Bag X1, Matieland, 7602, South Africa; E-mail:

Submitted for publication March 2003.

Accepted for publication July 2003.

Muscle contraction is dependent on a cascade of events culminating in actomyosin cross-bridge interactions. Under activated conditions, actin and myosin interact in a cyclical fashion, utilizing adenosine triphosphate (ATP) and producing force (8). Some of the properties of active muscle that are frequently described are its maximal capacities for force production (Po) (16), unloaded shortening velocity (Vmax) (41), and the actomyosin ATPase rate (42). The relationship between the mechanical and chemical properties of active muscle can be described as the energy cost or metabolic cost of the work done or the isometric tension maintained (30). This chemomechanical property is frequently termed efficiency (2,17). Mechanical, chemical, and chemomechanical properties vary for different muscles with different characteristic fiber types in small mammals (4) and for different fibers, even from the same muscle of human subjects (27). The heterogenous properties of skeletal muscle and its capacity to express an appropriate form to suit the requirements of different conditions are proof of its long-term malleability and potential for acclimation. However, the main issues of this review are the changes in the mechanical and chemomechanical properties in response to a specific acute condition: accumulation of metabolites from high-intensity, relatively short-duration exercise to fatigue.

It is well known that an acute transition from rest to exercise is accompanied by a greatly increased demand for energy by the working muscle and that the muscle attempts to balance this demand (29) utilizing various biochemical ADP-phosphorylating pathways for ATP supply (49). Because ATP is the primary chemical energy source for the contractile cycle, the ideal situation for continuation of exercise is one of ATP homeostasis, where ATP supply indeed meets ATP demand (see Fig. 1) and there is a tight coupling between metabolic capacity and work output (28), indicated in Figure 1 by the matching block arrows. Current theories hold that the sensing of increased ATP demand is accomplished, at least in part, by feedback from metabolites that accumulate during exercise, whether they are the metabolites of ATP hydrolysis itself or metabolites of the various ATP supply pathways (39). Additional roles in feedback are those played by ionic calcium and redox potential (39), and a permissive role for oxygen has been proposed (29). These mechanisms (and others not yet elucidated) maintain ATP homeostasis with remarkable success under steady-state exercise conditions with a range of intensities (28). Although ATP utilization rate depends linearly on exercise intensity, other factors may more subtly modulate it, e.g., availability of oxygen (49), fiber type (4), duration of the work bout (30), and whether muscle is shortening or lengthening during contraction (43).



Fatigue during exercise is a physiological phenomenon whose different impacts have provoked much controversy in the field of exercise science. The mechanisms of fatigue are complex, and both central and peripheral components have been discussed at length before (19,21,54) and do not form the focus of this paper. However, in the final section, I will propose that a link between these apparently opposing mechanisms may lie at the cross-bridges themselves. At this point, it is sufficient to mention that there is certainly consensus that high exercise intensities, demanding high rates of cross-bridge turnover and ATP utilization, can substantially change the intracellular milieu of the working muscle (6,14,32). This is indicated in Figure 2 by the absence of homeostasis and greater accumulation of metabolites compared with Figure 1. Despite very high ATP utilization rates during fatiguing exercise, ATP concentration ([ATP]) itself is fairly well defended (28), and only moderate reductions in [ATP] of around 30% have been seen in biochemical assessments of homogenates from muscle biopsies taken at fatigue (6). Even less change is seen in experiments of intact muscle using 31P NMR (38). In the face of any imbalance between the demand and subsequent supply of ATP, it would make teleological sense for some short-term regulation to occur at the sites of ATP utilization themselves in an attempt to reduce the imbalance. Possible solutions, as indicated in Figure 2, include downregulation of the demand for ATP or an increase in the efficiency with which the cross-bridges utilize that ATP. This paper will focus on one particular aspect of fatigue: intramuscular accumulation of metabolites of ATP hydrolysis to such an extent that they significantly influence the myofibrillar contractile cycle (indicated by dashed arrows in Fig. 2).



The mechanical manifestations of fatigue include a decline in force production, a reduced velocity of contraction, and a flattening of the force-velocity curve (21). In addition, there is evidence for improved efficiency of contraction in human subjects (3). The mechanism for the improved efficiency is not fully elucidated, and the role of metabolites of ATP hydrolysis will be addressed in this review. Specifically, the role of ADP will be highlighted because the roles of phosphate and hydrogen ions have received the most focus in the past (e.g., 10,35,43,53). Our current concept of the extent of [ATP] reduction and [ADP] accumulation during fatigue will be questioned (see section Metabolite Accumulation in Vivo) in the light of evidence from single-fiber biochemical analyses (32) and other models (36,37).

Back to Top | Article Outline


The efficiency of muscle contraction is defined as the proportion of total energy utilization that is used specifically for mechanical work output and is expressed as a ratio or as a percentage of total. Different terms have been used for this property of muscle, such as thermodynamic efficiency (27), efficiency of work performance (17), mechanical efficiency (4), and metabolic energy cost of steady state force (30). Some of these models are briefly elucidated below. However, first, let us consider the components of “muscle efficiency.” In muscle, ATP is supplied by metabolic pathways and is utilized primarily by the actomyosin ATPases, the Na+/K+ ATPases, and the sarcoplasmic reticulum Ca2+ ATPases. Any other energy utilized is dissipated as heat. Only the cross-bridges produce force that is translated externally to mechanical work. Therefore, total energy change during contraction is equivalent to the sum of the mechanical output and the heat output (4). This has been called myothermal energy balance (38). For various reasons, the two sides of the myothermal energy balance are not easy to measure experimentally and many different indirect models of assessment exist (2).

During steady-state whole-body exercise, ATP supplied for work is determined indirectly from oxygen consumption, and the work done is measured externally depending on the exercise device utilized. With this approach, the work output of all the muscles involved is not measured, and the energy dissipated as heat is not determined. Hence, a true efficiency of muscle contraction cannot be calculated but rather a gross efficiency for performance. Frequently, the energy utilized by the whole body during the specified exercise is reported. During running at submaximal intensity, it is assumed that all the energy is provided aerobically and the substitute for muscle efficiency is described as “running economy” at a fixed, steady-state speed (55). Similarly, for cycle ergometer exercise, a “gross efficiency” is determined from the ratio of work accomplished during pedaling to calculated caloric expenditure over a specified time (11). In the latter example, true mechanical efficiency of muscular work is also not determined because some of the measured energy expenditure is not involved directly in pedaling. Despite the simplicity of these in vivo models, as soon as the exercise changes, even subtly, the economy may also change. For example, when cycling on an ergometer at a fixed workload, the oxygen cost goes up if pedal frequency is increased (22).

To more closely estimate the efficiency of contracting muscle in human subjects, it is necessary to use a model with a clearly defined and quantifiable mechanical output of a specific working muscle (or muscle group) and a method to more accurately determine the total energy utilization or conversion of the same specific muscle group. As reviewed in detail by Bangsbo (2), several different approaches have been used depending on the intensity of the work and whether both aerobic and anaerobic energy sources are utilized, or aerobic alone. Ryschon et al. (46) used 31P NMR to indirectly determine a rate of ATP utilization, whereas others have measured the oxygen consumption across the specific muscle groups participating in the measured work as well as estimating heat production and the contribution of anaerobic energy production (20,24). These and other in vivo experiments (2) indicate that: i) the efficiency of concentric exercise is less than that of eccentric exercise, which is less than that of isometric exercise at a given external workload (46); ii) dynamic contractions with a lower load and higher frequency are less efficient than higher load, less frequent contractions despite equivalent total work output (20); and iii) mechanical efficiency of the quadriceps is higher at the end of exercise to fatigue than at the beginning (2), although when exercise is not done to fatigue, efficiency becomes less during the later stages of the exercise bout (3). The enhanced efficiency during fatigue occurs because the decrement in ATP utilization during fatiguing exercise is greater than the decrement in work output (2), whereas when work output remains constant throughout, ATP turnover increases with duration (3). This paper will now discuss some of the possible mechanisms for, specifically, the changes in efficiency with fatigue.

In attempts to elucidate the mechanisms, many muscle physiologists have used small animal models and tested whole, but isolated, muscle mechanics and oxygen consumption in situ (16) or even in small bundles of muscle fibers in vitro (30). Similar to results in human subjects, when single contractions of varying durations were assessed in vitro, the longer contractions were more economical (30). A similar result was achieved when de Haan et al. (15) calculated energy utilization from biochemical alterations in the muscle instead of from oxygen consumption. Efficiency of contraction is frequently determined by assessing total energy output rather than energy utilization. Efficiency is then calculated as the ratio between mechanical output and the sum of the mechanical and heat outputs (4,5). Not surprisingly, given the diversity inherent in skeletal muscle, some of these experiments have shown that efficiency is a characteristic that varies between muscles of differing fiber types (4). During isometric tetanic contractions, mouse soleus produced five times less heat than extensor digitorum longus per unit of force (4). When calculating the energetics of contraction with their model, Barclay et al. (5) are also able to partition the cross-bridge and noncross-bridge portions of energy utilization, thus enabling a calculation of mechanical efficiency of the cross-bridges themselves. In an experiment investigating the effects of fatigue on efficiency, their data indicated a decline in efficiency of the noncross-bridge component but an improvement of efficiency when considering the cross-bridge component. The latter can be interpreted from the data showing that force declined by only 7.5%, whereas cross-bridge related heat declined by 17% from the start to the end of the fatiguing series of tetani. Whereas in whole muscle a variety of processes individually, or in concert, affect efficiency (24) and overall efficiency is not improved by fatigue in some models (3), the skinned fiber model probes only the cross-bridge cycle (42). Therefore, this model can probe changes in efficiency at a single site.

In the chemically skinned muscle fiber model, both the sarcolemmal ATPases and sarcoplasmic reticulum ATPases are rendered inactive by the skinning procedure. Hence any determination of ATP utilization is from actomyosin ATPases only. In this model, force and ATP utilization are not always measured over a specified time period, as in the previous models discussed above, but at a particular point in time and for a certain set of conditions (42). The resultant chemomechanical property is termed “tension economy.” Also, because there is a good correlation between actomyosin ATPase rate and unloaded shortening velocity (45), an indirect indication of tension economy can even be assessed in mechanical experiments alone. However, even in the skinned fiber model, the mode of contraction may differ, particularly with respect to isometric versus shortening contractions. For the rest of this review, only the effects of metabolites of fatigue on skinned fibers undergoing isometric contraction will be considered.

Back to Top | Article Outline


During fatigue multiple metabolic changes occur simultaneously. The beauty of the in vitro skinned fiber model is that it gives researchers easy access to and control of intracellular conditions. Another advantage (depending on the purpose of the experiment) is that the cross-bridges can be evaluated without the contributions of neural activity and sarcolemmal and sarcoplasmic reticulum functions. A final advantage is that, as opposed to energetic assessments of actin and myosin interaction in solution, the myofibrillar array is intact. Single components of the bathing medium can be manipulated to simulate selected conditions of fatigue to enable investigators to determine the specific effects of the manipulated component rather than the condition of fatigue as a whole.

Back to Top | Article Outline

Phosphate and hydrogen ions.

It is well known from in vivo experiments using 31P NMR that inorganic phosphate (Pi) accumulates during fatigue in human subjects (14) and in smaller mammals (38). In the 1980s, several experiments on single skinned fibers from mammalian fast-twitch muscle proved that adding Pi to the bathing solution indeed caused a decrease in maximal isometric tension production (9). This direct effect on tension is linearly related to the log of the concentration of Pi (9) and hence would be greatly enhanced at exhaustion from high-intensity exercise compared with fatigue from more moderate-intensity exercise concomitant with lesser perturbations in Pi.

The relevance of these in vitro experiments to conditions during in vivo exercise to fatigue can be seen when one considers that the accumulation of Pi in vivo mirrors the decline in PCr, both of which may be observed using 31P NMR (38). Also, de Ruiter et al. (16) have shown in situ that the full recovery of force during isometric contractions of rat soleus muscle after fatigue is very closely associated with the restoration of phosphocreatine (PCr). Recent evidence suggests that the effect of Pi is less apparent at higher temperatures than at 5°C; however, the effect on force is still substantial even at 30°C (10). In contrast, neither the maximal velocity of unloaded shortening (9) nor the maximal actomyosin ATPase rate (9,35) are much affected by Pi accumulation, although not all authors concur (43). Nevertheless, even the reduction in ATPase rate of just over 20% reported by Potma and Stienen (43) with addition of 30 mM Pi, is less than the reduction in force at the same concentration of Pi. Hence, Pi accumulation cannot directly explain all phenomena that have been observed during fatigue. For example, in the classic study of Crow and Kushmerick (12), fatigue was associated with an approximately 50% decrease in velocity of contraction in fast-twitch mouse muscle. Similar reductions were shown in cat fast-twitch, fatigue-resistant muscle (26). Reduced ATP utilization rates have been ascribed to hydrogen ion (H+) accumulation (21), but it is now clear that the effect of H+ on contraction velocity is not duplicated when experiments take place at higher temperatures (∼30°C) in skinned fibers, intact fibers, or whole muscles in situ (53). However, as will become evident at the end of this review, it is important to note that temperature influences do not change the essential conclusions on the effects of Pi on contractile mechanics (16,43): its major effect is to reduce force.

Back to Top | Article Outline

Adenosine diphosphate.

Because adenosine diphosphate (ADP) is also a direct product of ATP hydrolysis, its effects on the contractile cycle and myofibrillar ATPase activity have also been considered by several investigators (see Table 1). The conditions and findings in these experiments were quite inconsistent. In summary, Cooke and Pate (9), Metzger (41), and Myburgh and Cooke (42) demonstrated quite a large effect of added ADP on isometric tension (an increase of 15% or more) and cross-bridge cycling rate (a decrease of between 27 and 52%), which other authors did not show, albeit under different conditions. The models of the cross-bridge cycle predict that ADP should compete with ATP for the nucleotide binding site and that accumulation of ADP would slow down cross-bridge detachment, ATPase rate, and Vmax and concurrently potentiate force production (13). Although these effects have been shown in some studies, the concentrations of ADP in the studies were unphysiologically high (Table 1). Whether ADP accumulation plays a role under physiological conditions remains controversial, with authors considering its effect small (23) or insignificant (7,34) (Table 1).



All the above experiments were possibly influenced by experimental complications such as gradients in the concentrations of various nucleotides between the bathing medium and the center of the fibers (7), the absence of an ATP regenerating system (9), or unphysiologically high concentrations of other components of the bathing medium (7,34). Karatzaferi et al. (33) recently investigated the effects of ADP on contraction in a skinned fiber model using a spin-labeled ADP analog (SL-ADP) that has a very low affinity for creatine kinase and is not easily phosphorylated. Its concentration can therefore be much more accurately controlled, and creatine kinase can be added for ATP regeneration. Other advantages of this system are the facts that: i) all other components of the bathing medium could be maintained despite the addition of the SL-ADP; ii) experiments could even be done at very low concentrations of ATP without any influence of the added SL-ADP on the ATP concentration; and iii) the affinity with which the analog binds to the actomyosin complex is equal to that of ADP itself, enabling both the competition between ATP and ADP to be fully elucidated and the reversibility of any finding to be checked on return to a baseline condition. Of relevance to the discussion here are the insights, elucidated by these highly controlled experimental conditions, related to the effects of ADP on the economy of contraction.

A consistent finding of Karatzaferi et al. (33) was an increase in isometric tension with addition of a range of SL-ADP concentrations that ranged from 100 μM to 2 mM. The addition of SL-ADP across this range was done under various conditions in which other components of the “intracellular milieu” (bathing medium) had been altered. For example, [ATP] was varied between 50, 150, and 500 μM; [Pi] was varied between 3, 10, and 54 mM; and pH was either pH 7 or pH 6. The improvement in tension with added SL-ADP was determined for each condition, and despite an approximately 10-fold difference in Po between differing baseline conditions, the absolute increments in P differed much less, indicating a comparatively consistent absolute effect of ADP. But the quantity of SL-ADP required to compete effectively with ATP under these different conditions varied considerably and ranged between only 100 μM SL-ADP when [ATP] was extremely low (50 μM), to 1.63 mM of SL-ADP when [ATP] was 0.5 mM. These data indicate that an effect of physiological quantities of ADP on tension is dependent on the existence of low concentrations of ATP. However, the relative improvement in Po (expressed as a percentage of the baseline tension) varied greatly depending on the conditions and was greatest under conditions when baseline Po was low (see Fig. 3) (33). These data indicate that ADP as an alleviator of fatigue may be relatively important especially when low pH and high [Pi] are detrimental to force production. In experiments determining the isometric ATPase activity, Karatzaferi et al. (33) found that SL-ADP was a competitive inhibitor of ATP binding and indeed slowed down the ATPase rate. The effect of SL-ADP was half maximal at a concentration of 240 μM, indicating that an effect could be seen within the range of physiological accumulation of ADP. However, the affinity of ATP for the nucleotide binding site is greater than that of SL-ADP, as indicated by the low Km for ATP of close to 30 μM. Decreasing the pH of the bathing solution did not significantly alter the effective concentration of SL-ADP but enhanced the affinity of ATP for the binding site. Despite the theories implicating ADP release as the rate limiting step in the cross-bridge cycle (9,13), these data indicate that during fatigue a slowing down of the ATPase rate may be effected by ADP only when there is also a substantial reduction in [ATP].



Back to Top | Article Outline


The question that arises now, is whether the conditions under which ADP may play a significant role in alleviating fatigue occur in vivo. As mentioned before, and as reviewed by Fitts (21), the general consensus of 31P NMR studies is that [PCr] can be reduced to 5% of resting levels within 30 s of maximal sprint exercise, whereas the reduction in [ATP] is modest and certainly not within the range suggested to be significant for mechanics (23). Analysis of homogenates of muscle biopsy samples have concurred, but in single fibers dissected from muscle biopsies taken at fatigue (after electrical stimulation with occluded blood flow), 11% of the fibers could have [ATP] lower than 40% of the resting value (48). In a more recent study (32), after 25 s of maximal voluntary cycle sprinting, the reported means for [ATP] in single fibers from different fiber types were below 40% of resting values in all cases. In addition, in this study the mean [ATP] of those fibers containing any fraction of Type IIX myosin heavy chain (Type IIAx and IIXa hybrids) was only 20% of resting concentrations (5 ± 3 μmol·g−1 d.m.). Although these single fiber data give a clearer picture of the extent to which [ATP] can decline in those fibers with the highest myosin ATPase rates, it must still be taken into account that there is a delay of a few seconds after cessation of exercise before freezing of biopsy samples, during which time [ATP] may already be recovering. Also, microenvironments may exist within the fibers (37) where the imbalance between ATP supply and demand is greater than the average for the cell. A theoretical model of Kemp et al. (36) suggests that spatial differences in adenine nucleotide concentrations within muscle in response to increases in ATPase rates are in the direction of higher [ATP] at the mitochondria rather than at the ATPase sites and higher [ADP] at the ATPase sites rather than the mitochondria. It is possible that a new technological advance, the monitoring of intracellular [ATP] in contracting muscle fibers in vitro using a firefly luciferin/luciferase reaction (1), may in the future elucidate spatial dissimilarities in [ATP] with extreme fatigue. Furthermore, calculated maximal [ADP] is based on the assumption that [ATP] is evenly distributed in the cell and that the creatine kinase reaction is in equilibrium, but Kemp et al. (36) suggest that the creatine kinase reaction may not necessarily be in equilibrium in vivo. Therefore, true ADP concentrations may have been underestimated in the past. The recent paper by Karatzaferi et al. (33) indicates for the first time in detail the extent to which ADP would need to be present relative to the extent of ATP decline to have functionally significant effects on cross-bridge mechanics.

A study by Metzger (41) indicated in skinned fibers that two separate “populations” of cross-bridges occur under conditions of added ADP (5 mM): the one corresponding to those seen by Cooke and Pate (9) and the other substantially slower. This observation could possibly be explained if ADP rebound to the actomyosin nucleotide pocket after its release and this conformation of actomyosin-ADP was different and inhibited shortening velocity more than the “usual” isomer of actomyosin-ADP occurring prior to release of ADP after the power stroke. The possibility that the cross-bridge cycle could be reversed at this late stage in the cycle is posed in a model in Karatzaferi et al. (33). This model fitted the SL-ADP data from actual experiments very well. Although this is still a theoretical concept, it is probably safe to assume that if reversal occurred, it would occur more frequently and slow down the cross-bridge cycle more under conditions of ADP accumulation. As reviewed by Korge and Campbell (37), several papers have discussed the possibility that specifically ADP diffusion may be less than optimal and therefore that functionally significant accumulation of [ADP] may occur during fatigue, particularly at sites with the highest ATPase rates.

Back to Top | Article Outline


Adjustments in ATP supply or modulation of ATP utilization and efficiency at the sites of the cross-bridges themselves represent appropriate regulatory mechanisms developed by skeletal muscle to match ATP demand and supply. However, these solutions are not always adequate, and in order to defend intracellular [ATP] and to prevent rigor-like cross-bridges (that could theoretically occur very rapidly given the high ATP turnover rates during exercise), conditions of mismatching must be rapidly sensed and communicated and a rapid response must be elicited (see Fig. 4). Ultimately, the best solution is to substantially downregulate ATP demand by inhibition of activation.



Crucial to activation of muscle contraction is the excitation-contraction coupling process that involves many steps, each of which could potentially be affected by metabolite accumulation (for recent review see Stephenson et al. (50)). Accumulation of ions such as Mg2+, Ca2+, H+, or inorganic phosphate may affect the ryanodine receptor/calcium release channels, thus reducing activation. Particularly relevant to the prior discussion in this review, accumulation of Mg2+ in the mM range alongside significant declines in [ATP] in microenvironments surrounding the ryanodine receptor may decrease the number of open Ca2+ release channels in response to depolarization (50). Furthermore, the duration that the Ca2+ release channels remain open is reduced by a reduction in free calcium in the lumen of the sarcoplasmic reticulum (47), a scenario that can occur late in fatigue if inorganic phosphate enters the sarcoplasmic reticulum lumen (52). A third mechanism for peripheral reduction of activation by impaired excitation-contraction coupling is the actual binding of Ca2+ to the regulatory filaments, a process that may be altered during fatigue, thus impairing the number of active actomyosin cross-bridges (for review see ref. 21).

Nevertheless, proponents of a central cause for fatigue cite much evidence for a central inhibition of neural activation (19). A different interpretation is that decreased muscle activation is not a centrally controlled cause of fatigue but rather a solution to existing peripheral fatigue. Various localized metabolic sensors are known to exist that can have far-reaching consequences, e.g., the hypothalamus’s glucose-sensitive neurons (18), the pancreas’s substrate-sensitive mitochondria (18), and the muscle’s metaboreflex (51) and AMP sensor (25). Some of these sensors interact with or are part of the nervous system, whereas others act intracellularly. During high-intensity exercise to fatigue, AMP-kinase is teleologically the most likely sensor to be operative in muscle, but direct effects on the cross-bridge cycle, the sarcoplasmic reticulum, or communication with the nervous system remain to be demonstrated.

An alternative explanation is that sensors for peripheral fatigue other than metabolic sensors may exist, e.g., a mechanosensor. I speculate that a mechanosensor’s feedback of actual force production may play a role in modulating ATP demand. When there is a mismatch between the expected force production and the peripheral reality fed back by the putative mechanosensor, the result may be a rapid decrease in activation of the working muscle that can ultimately have a large effect on decreasing ATP demand so as to better match ATP supply (see Fig. 4) even before fatigue. The Golgi tendon organ (GTO) is well known to provide feedback regarding muscle force production, although some uncertainties still exist (44). The role of the GTO in fatigue is not well studied, but its afferent activity is altered (31). I hypothesize that this is in part in response to the well-known effect of phosphate accumulation on cross-bridge force production. This hypothesis implies that an early step in the link between the apparently opposing peripheral and central mechanisms of fatigue may lie in a localized effect at the cross-bridges themselves and that phosphate may be responsible, at least in part, for alleviation of fatigue.

Back to Top | Article Outline


This paper elucidates a positive role for metabolites in skeletal muscle metabolism and fatigue. Metabolites communicate that the ATP demand has increased, so that the ATP supply can be upregulated to meet the demand. Under conditions of fatigue, metabolites can downregulate the myosin ATPase demand by slowing down the cross-bridge cycle. When fatigue is more severe and ADP accumulates to such an extent that it competes effectively with ATP for the nucleotide binding site, force is potentiated and ATPase rate is reduced, so that cross-bridge efficiency is improved thus at least partially downregulating the muscle’s overall ATP demand per unit of force output. Finally, this paper proposes that the link between peripheral fatigue and central fatigue is also a result of metabolite accumulation, particularly Pi. This proposal depends on the speculation that central awareness of mismatches between expected force and produced force are sensed peripherally, and neural activation is ultimately downregulated in addition to local downregulation of activation due to altered excitation-contraction coupling.

I would like to thank Roger Cooke for first challenging me to try to solve the issue of enhanced tension economy in skeletal muscle during fatigue when I was a Postdoctoral Fellow in his laboratory and for the many subsequent opportunities to contribute to work on this issue. A special word of thanks to Christina Karatzaferi for her hard work in expanding and completing the most recent work (33) and to Peter Hochachka for discussions on downregulation of ATP demand.

Back to Top | Article Outline


1. Allen, D. G., J. Lännergren, and H. Westerblad. Intracellular ATP measured with luciferin/luciferase in isolated mouse skeletal muscle fibres. Pflügers Arch. 443: 836–842, 2002.
2. Bangsbo, J. Physiological factors associated with efficiency in high intensity exercise. Sports Med. 22: 299–306, 1996.
3. Bangsbo, J., P. Krustrup, J. Gonzalez-Alonso, and B. Saltin. ATP production and efficiency of human skeletal muscle during intense exercise: effect of previous exercise. Am. J. Physiol. Endocrinol. Metab. 280: E956–E964, 2001.
4. Barclay, C. J., J. K. Constable, and C. L. Gibbs. Energetics of fast- and slow-twitch muscles of the mouse. J. Physiol. 472: 61–80, 1993.
5. Barclay, C. J., N. A. Curtin, and R. C. Woledge. Changes in crossbridge and non-crossbridge energetics during moderate fatigue of frog muscle fibres. J. Physiol. 468: 543–555, 1993.
6. Bogdanis, G. C., M. E. Nevill, H. K. A. Lakomy, and L. H. Boobis. Power output and muscle metabolism during and following recovery from 10 and 20 s of maximal sprint exercise in humans. Acta Physiol. Scand. 163: 261–272, 1998.
7. Chase, P. B., and M. J. Kushmerick. Effect of physiological ADP concentrations on contraction of single skinned fibers from rabbit fast and slow muscles. Am. J. Physiol. 268: C480–C489, 1995.
8. Cooke, R. Actomyosin interaction in striated muscle. Physiol. Rev. 77: 671–697, 1997.
9. Cooke, R., and E. Pate. The effects of ADP and phosphate on the contraction of muscle fibers. Biophys. J. 48: 789–798, 1985.
10. Coupland, M. E., E. Puchert, and K. W. Ranatunga. Temperature dependence of active tension in mammalian (rabbit psoas) muscle fibres: effect of inorganic phosphate. J. Physiol. 536: 879–891, 2001.
11. Coyle, E. F., L. S. Sidossis, J. F. Horowitz, and J. D. Beltz. Cycling efficiency is related to the percentage of type I muscle fibres. Med. Sci. Sports Exerc. 24: 782–788, 1992.
12. Crow, M. T., and M. J. Kushmerick. Chemical energetics of slow- and fast-twitch muscles of the mouse. J. Gen. Physiol. 79: 147–709, 1982.
13. Dantzig, J. A., M. G. Hibberd, D. R. Trentham, and Y. E. Goldman. Cross-bridge kinetics in the presence of MgADP investigated by photolysis of caged ATP in rabbit psoas muscle fibres. J. Physiol. 432: 639–680, 1991.
14. Dawson, M. J., D. G. Gadian, and D. R. Wilkie. Muscular fatigue investigated by phosphorous nuclear magnetic resonance. Nature. 274: 861–866, 1978.
15. De Haan, A., J.De Jong, J. E. Van Doorn, P. A. Huijing, R. D. Woittiez, and H. G. Westra. Muscle economy of isometric contractions as a function of stimulation time and relative muscle length. Pflügers Arch. 407: 445–450, 1986.
16. De Ruiter, C. J., A. De Haan, and A. J. Sargeant. Repeated force production and metabolites in two medial gastrocnemius muscle compartments of the rat. J. Appl. Physiol. 79: 1855–1861, 1995.
17. Di Prampero, P. E., U. Boutellier, and A. Marguearat. Efficiency of work performance and contraction velocity in isotonic tetani of frog sartorius. Pflügers Arch. 412: 455–461, 1988.
18. Duchen, M. R. Contributions of mitochondria to animal physiology: from homeostatic sensor to calcium signaling and cell death. J. Physiol. 516: 1–17, 1999.
19. Enoka, R. M., and D. G. Stuart. Neurobiology of muscle fatigue. J. Appl. Physiol. 72: 1631–1648, 1992.
20. Ferguson, R. A., D. Ball, P. Krustrup, et al. Muscle oxygen uptake and energy turnover during dynamic exercise at different contraction frequencies in humans. J. Physiol. 536: 261–271, 2001.
21. Fitts, R. H. Cellular mechanisms of muscle fatigue. Physiol. Rev. 74: 49–94, 1994.
22. Gaesser, G. A., and G. A. Brooks. Muscular efficiency during steady-rate exercise: effects of speed and work rate. J. Appl. Physiol. 38: 1132–1139, 1975.
23. Godt, R. E., and T. M. Nosek. Changes of intracellular milieu with fatigue or hypoxia depress contraction of skinned rabbit skeletal and cardiac muscle. J. Physiol. 412: 155–180, 1989.
24. Gonzalez-Alonso, J., B. Quistorff, P. Krustrup, J. Bangsbo, and B. Saltin. Heat production in human skeletal muscle at the onset of intense dynamic exercise. J. Physiol. 524: 603–615, 2000.
25. Hardie, D. G., and D. Carling. The AMP-activated protein kinase - fuel gauge of the mammalian cell? Eur. J. Biochem. 246: 259–273, 1997.
26. Hatcher, D. D., and A. R. Luff. Force-velocity properties of fatigue-resistant units in cat fast-twitch muscle after fatigue. J. Appl. Physiol. 63: 1511–1518, 1987.
27. He, Z.-H., R. Bottinelli, M. A. Pellegrino, M. A. Ferenczi, and C. Reggiani. ATP consumption and efficiency of human single muscle fibers with different myosin isoform composition. Biophys. J. 79: 945–961, 2000.
28. Hochachka, P. W., and G. O. Matheson. Regulating ATP turnover rates over broad dynamic work ranges in skeletal muscles. J. Appl. Physiol. 73: 1697–1703, 1992.
29. Hochachka, P. W., and G. B. McClelland. Cellular metabolic homeostasis during large-scale change in ATP turnover rates in muscles. J. Exp. Biol. 200: 381–386, 1997.
30. Hogan, M. C., E. Inghan, and S. Sadi Kurdak. Contraction duration affects metabolic energy cost and fatigue in skeletal muscle. Am. J. Physiol. Endocrinol. Metab. 274: E397–E402, 1998.
31. Hutton, R. S., and L. D. Nelson. Stretch sensitivity of golgi tendon organs in fatigued gastrocnemius muscle. Med. Sci. Sports Exerc. 18: 69–74, 1986.
32. Karatzaferi, C., A. de Haan, R. A. Ferguson, W. Van Mechelen, and A. J. Sargeant. Phosphocreatine and ATP content in human single muscle fibres before and after maximal dynamic exercise. Pflügers Arch. 442: 467–474, 2001.
33. Karatzaferi, C., K. H. Myburgh, M. K. Chinn, K. Franks-Skiba, and R. Cooke. Effect of an ADP analog on isometric force and ATPase activity of active muscle fibers. Am. J. Physiol. Cell Physiol. 284: C816–C825, 2003.
34. Kawai, M. K., and H. R. Halvorson. Role of MgATP and MgADP in the cross-bridge kinetics in chemically skinned rabbit psoas fibers. Biophys. J. 55: 595–603, 1989.
35. Kawai, M. K., K. Güth, K. Winnikes, C. Haist, and J. C. Rüegg. The effect of inorganic phosphate on the ATP hydrolysis rate and the tension transients in chemically skinned rabbit psoas fibers. Pflügers Arch. 408: 1–9, 1987.
36. Kemp, G. J., D. N. Manners, F. J. Clark, M. E. Bastin, and G. K. Radda. Theoretical modelling of some spatial and temporal aspects of the mitochondrion/creatine kinase/myofibril system in muscle. Mol. Cell. Biochem. 184: 249–89, 1998.
37. Korge, P., and K. B. Campbell. The importance of ATPase micro-environment in muscle fatigue: a hypothesis. Int. J. Sports Med. 16: 172–179, 1995.
38. Kushmerick, M. J., and R. A. Meyer. Chemical changes in rat leg muscle by phosphorus nuclear magnetic resonance. Am. J. Physiol. Cell. Physiol. 248: C542–C549, 1985.
39. Meyer, R. A., and J. M. Foley. Cellular processes integrating the metabolic response to exercise. In: Handbook of Physiology, Section 12, Exercise: Regulation and Integration of Multiple Systems, L. B. Rowelland J. T. Shepherd (Eds.). 1996, pp. 841–869.
40. Lu, Z., R. L. Moss, and J. W. Walker. Tension transients initiated by photogeneration of MgADP in skinned skeletal muscle fibers. J. Gen. Physiol. 101: 867–88, 1993.
41. Metzger, J. M. Effects of phosphate and ADP on shortening velocity during maximal and submaximal calcium activation of the thin filament in skeletal muscle fibres. Biophys. J. 70: 409–417, 1996.
42. Myburgh, K. H., and R. Cooke. Response of compressed skinned skeletal muscle fibers to conditions that simulate fatigue. J. Appl. Physiol. 82: 1297–1304, 1997.
43. Potma, E. J., and G. J. M. Stienen. Increase in ATP consumption during shortening in skinned fibres from rabbit psoas muscle: effects of inorganic phosphate. J. Physiol. 496: 1–12, 1996.
44. Proske, U., and J. E. Gregory. Signalling properties of muscle spindles and tendon organs. Adv. Exp. Med. Biol. 508: 5–12, 2002.
45. Reggiani, C., E. J. Potma, R. Bottinelli, M. Canepari, M. A. Pellegrino, and G. J. M. Stienen. Chemo-mechanical energy transduction in relation to myosin isoform composition in skeletal muscle fibres of the rat. J. Physiol. 502: 449–460.
46. Ryschon, T. W., M. D. Fowler, R. E. Wysong, A-R. Anthony, and R. S. Balaban. Efficiency of human skeletal muscle in vivo: comparison of isometric, concentric, and eccentric muscle action. J. Appl. Physiol. 83: 867–874, 1997.
47. Sitsapesan, R., and A. J. Williams. The gating of the sheep skeletal sarcoplasmic reticulum Ca2+-release channel is regulated by luminal Ca2+. J. Membr. Biol. 146: 133–144, 1995.
48. Söderlund, K., and E. Hultman. ATP content in single fibres from human skeletal muscle after electrical stimulation and during recovery. Acta Physiol. Scand. 139: 459–466, 1990.
49. Stary, C. M., and M. C. Hogan. Phosphorylating pathways and fatigue development in contracting Xenopus single skeletal muscle fibers. Am. J. Physiol. Regul. Integr. Comp. Physiol. 278: R587–R591, 2000.
50. Stephenson, D. G., G. D. Lamb, and G. M. M. Stephenson. Events of the excitation-contraction-relaxation (E-C-R) cycle in fast- and slow-twitch mammalian muscle fibres relevant to muscle fatigue. Acta Physiol. Scand. 162: 229–245, 1998.
51. Vissing, J., D. A. MacLean, M. Sander, B. Saltin, and R. G. Haller. The exercise metaboreflex is maintained in the absence of muscle acidosis: insights from muscle microdialysis in humans with McArdle’s disease. J. Physiol. 537: 641–649, 2001.
52. Westerblad, H., and D. G. Allen. The effects of intracellular injections of phosphate on intracellular calcium and force in single fibres of mouse skeletal muscle. Pflügers Arch. 431: 964–970, 1996.
53. Westerblad, H., D. G. Allen, and J. Lännergren. Muscle fatigue: lactic acid or inorganic phosphate the major cause? News Physiol. Sci. 17: 17–21, 2002.
54. Westerblad, H., J. A. Lee, J. Lännergren, and D. G. Allen. Cellular mechanisms of fatigue in skeletal muscle. Am. J. Physiol. Cell. Physiol. 261: C195–C209, 1991.
55. Weston, A. R., Z. Mbambo, and K. H. Myburgh. Running economy of African and Caucasian distance runners. Med. Sci. Sports Exerc. 32: 1130–1134, 2000.


©2004The American College of Sports Medicine